Fix flaky test in TestClassificationEnsembles by stabilizing HashMap iteration order #420
Conversation
…iteration order Signed-off-by: lbh930 <[email protected]>
|
Thank you for your pull request and welcome to our community! To contribute, please sign the Oracle Contributor Agreement (OCA).
To sign the OCA, please create an Oracle account and sign the OCA in Oracle's Contributor Agreement Application. When signing the OCA, please provide your GitHub username. After signing the OCA and getting an OCA approval from Oracle, this PR will be automatically updated. If you are an Oracle employee, please make sure that you are a member of the main Oracle GitHub organization, and your membership in this organization is public. |
oracle-contributor-agreement
bot
added
the
OCA Required
|
What's the stack trace you get from a test failure here? The features map isn't probed by the test methods, just tested that it exists and is not empty. |
|
Thank you for signing the OCA. |
oracle-contributor-agreement
bot
added
OCA Verified
|
Running nondex on this test gives me a different error which looks like it's something to do with the same kind of issue in the CART implementation (i.e. in AdaboostTrainer there are multiple locations which compute the feature id maps and these end up out of sync running under nondex). Unfortunately I only have a partial fix for the root cause it at the moment, so I need to work on it some more. |
|
Should this be closed in favour of #424? |
2 participants
Description
This PR fixes non-deterministic test flakiness in TestClassificationEnsembles by stabilizing HashMap iteration order during feature map assertions. The fix adds a sortFeatures() helper method to sort feature maps before comparison.
Motivation
This change is necessary to eliminate test flakiness caused by HashMap's non-deterministic iteration order in feature maps returned by getTopFeatures(). This is a bug fix that ensures test stability without modifying any production code. The fix is verified using NonDex.
Paper reference
N/A